As explained in the introduction, “viral marketing” is a modern term for the age-old concept of word-of-mouth advertising. The idea is that a product proliferates by people “spreading the word.” It is called “viral” because it can cause the product to spread like a virus, or “infect” its market.

     Customers and carriers. The “virus” in viral marketing may be a product, but more generally it is information about the product, which we will call here the “message.” It is the “word” in word-of-mouth advertising. Just as an individual can have a biological virus without having the associated disease, a person does not need to actually own or use a product to spread the word about it. As in biology, we can think of a “carrier” as someone who understands the products’s value proposition, but is for some reason not in a position to personally benefit from it. For example, a teenage boy excited about driving a Farrari can be a highly infectious agent, even though he is not able to either own or drive one. Sharing of bootleg software is another way that non-owners spread a product, and this is not always detrimental to business. The musical band, The Grateful Dead, made an artform of exploiting bootlegging to drive commercial popularity^«. Examples of people who spread the message about a product can include:

• Satisfied customers
• Wanna-be customers that can’t use or afford the product, but think it’s great for people who can
• Mavens, who like to demonstrate their knowledge and recommend what other people should do
• Journalists, including product reviewers, that check out a product and write about it
• For software products (including music and movies), pirates who obtain bootleg copies and then rave about it to their friends or online

     Virality. The effectiveness of a biological virus is the combination of its contagiousness (ability to spread) and virulence (ability to cause disease). There are viruses that live benignly in our bodies without causing any disease. These viruses may spread effectively, but their spreading has no effect on people having them. In effect, all infected individuals are carriers only. Such a case is disastrous in the marketing analog. It is not enough for the message to spread—it must also move people to action.

     We will use the word, “virality,” for the ability of a message to both spread and motivate sales. Virality depends on the characteristics and relationships of the message as well as its carriers, targets, and environment.

• The message. The first requirement is that the message be prolific, so that it tends to be repeated and reproduced often. In marketing, this comes from the value proposition repeatedly coming to the surface so that targets can be exposed to it. This can happen, in the case of active transmission (see below), by the message bearing enough significance in people’s minds and being memorable enough that it repeatedly comes to mind in conversations with different people.
• Carrier/target relationship. The message spreads more effectively if its carriers are connected and transmissive.
  • A connected carrier is one that has contact with a large number of targets.
  • A transmissive carrier is one who engages in activity that promotes transmission to targets.
• Message/environment relationship. The message must stand out against the cacophony of information competing every day for people’s attention and action. This is roughly equivalent to what Malcolm Gladwell calls “stickiness” in The Tipping Point. Some things that help make the message sticky are:
  • The associated product works really well, so the value proposition is strong.
  • The need and its solution can be communicated briefly in simple language, so the value proposition can be understood easily.
  • The message is embodied in an easy-to-remember phrase, picture, sound, or action. Better yet if it is also fun to repeat.
• Message/target relationship. For a biological virus to cause disease, it must penetrate the appropriate cell bodies. The analog in marketing is motivation for the target to use the product. The ability of a value proposition to motivate action depends on the depth of the underlying need, or relevance to the target. For example, a new formulation for a slimming diet may work very well and therefore have a strong value proposition. When a slender person hears about it, he may become a carrier and help to spread the word to overweight people that he knows, but the value proposition has little value to him directly and he will not likely become a customer himself.

     Role of carrier. In viral marketing, the carrier may take either an active or passive role in the transmission of the message. An active carrier is a person who talks or otherwise communicates directly with other people about the product. A passive carrier is someone who transmits the message inadvertently, even involuntarily. Passive transmission takes place when the message is conveyed by the product itself, such as the name emblazoned on a Tommy Hilfiger jersey, the hood ornament on a car, the tag line at the end of a HotMail message, and the greeting played to callers of Audix voice-mail customers.

     Cost. Viral marketing may be free or it may have direct and/or indirect costs associated with it.

• Free. If a product is so wonderful that people talk about it unprompted, then it may infect its market with no specific marketing effort on the part of the vendor. This can even take a vendor by surprise, as with the sudden reemergence of Hush Puppies as trendy footwear in the mid-1990s^« , or with the pre-launch popularity of EndNote software in 1988^«.
• Indirect cost. A vendor can actively engage its customers in recommending its products without providing any monetary reward for the favor. In this case, the cost is the time and materials used in the effort. This expense is usually miniscule compared to the cost of acquiring customers though any other advertising medium. Passive transmission has only indirect costs, such as the cost of a manufacturer nameplate on a car. Another example is the miniscule cost and effort when a realtor gives extra business cards to a satisfied seller and asks him to give them to homeowner friends.
• Direct cost. A vendor may also pay customers, either in cash or in goods and services, to make referrals. These payments may take the form of discounts on purchases, as in Amazon.com’s Share the Love campaign, a drawing for a more valuable prize, as in CallWave’s Tell Your Friends program, or direct commission on sales, as in a multi-level marketing program, such as Amway’s.

     While the term, “viral marketing”, relates to the biological phenomenon of viruses, the fundamental phenomenon of interest here is mathematical. It is the difference between geometric and exponential growth. Geometric growth takes place when a quantity changes by addition of some specific amount in successive steps. By contrast, in exponential growth, a quantity changes by a multiplicative factor with each step. Both types of growth apply to marketing and can be associated with different methods of attracting new business. In addition, growth due to viral transmission can be distinguished broken down into that done by new customers versus all customers. The effects are very different.

     Geometric growth is typically the result of broadcasting an advertising message. Here, “broadcasting” is used to mean any way of getting the message out to a large number of people, whether in print, “on the air,” or online. If an advertisement is aimed at generating specific business (as opposed to developing corporate image) and if it is done well, then one can generally predict and track the number and value of sales resulting from it. Furthermore, if increments to the advertising are planned and done just as well, then there will generally be a roughly linear relationship^« between the dollars spent and the business generated. If $100,000 worth of ads generates $1 million worth of business, then, if done right, spending another $100,000 may be expected to generate another $1 million. That is geometric growth. It works that way because each placement of an ad can be counted on to motivate a purchase reaction in a certain number of people. You spend a dollar to motivate a certain number of dollars of purchasing.

     The mathematics of broadcast advertising can be represented as follows. Suppose

• We start with N₀ customers on day 0.
• We spend $x per day on advertising with an effectiveness of acquiring q new customers per dollar spent.
• Each customer spends (or has a lifetime value of) $v.

N = N₀ + xqt

V = $ (xqtv)

     If the customer is more strongly motivated to talk about the product during an initial period after getting it, then we can model the growth in terms of the impact of new customers alone. Suppose

• We start with N₀ customers on day 0, m of whom are new (by some appropriate definition).
• Every new customer, on average, motivates n other people to become customers within d days.
• Each customer spends (or has a lifetime value of) $v.

N = N0 + m *

nt/d + 1 – n n – 1

V = $ m *

nt/d + 1 – n n – 1

* v

     If there is a level of enthusiasm or other motivation for talking about the product that can be maintained throughout ownership, or if passive transmission provides ongoing virality, then growth is modeled on the basis of leveraging the entire customer base. Suppose

• We start with N₀ customers on day 0.
• Every customer, on average, motivates n other people to become customers every d days.
• Each customer spends (or has a lifetime value of) $v.

N = N₀ (1 + n)^t/d

V = $ N₀ ( (1 + n)^t/d – 1) * v

     The individual effects of the three types of growth described above can be seen in the following graph.

Graph showing three types of growth independently. All scenarios assume starting with 1 million customers and cummulating growth effects every 30 days. Values for the individual parameters are selected to achieve a similar level of growth over a two-year period.

     The parameter values for the graph are chosen so that all three types of growth bring the company to double its customer base in about two years.

• Geometric (broadcast ads): $1,400 dollars a day is spent with an effectiveness of one customer acquired per dollar.
• Exponential due to new customers only: Starting with 25,000 new customers, each new customer brings in 1.04 others (or every 100 new customers bring in 104 more) after 30 days.
• Exponential due to all customers: Every customer brings in 0.03 new customers (or every 33^rd customer brings in one new one) after every 30 days.

     In this analysis, each of the three types of growth is considered independently. The calculation for broadcast ads assumes that there is no simulatneous viral effect going on. Similarly, the numbers for each type of viral growth assume that no additional new customers are being “bought” with advertising. Furthermore, the new customer scenario assumes that there is no continuing viral effect at all after someone has been a customer for 30 days, and the continuing viral scenario assumes that customers are no more inclined to talk about the product when they are new than at any later time.

     The results of the above analysis are fairly impressive for viral growth. However, they do not begin to show the actual power available from exponential effects. When two or more of the three effects are combined and allowed to interact, their mutual feedback creates truly astounding growth. For example, with the parameters used above to generate a doubling of the customer base in two years, when the effects are combined and allowed to interact, and even with a 10% churn rate factored in, the doubling occurs in less than five months! After two years with those parameters, if the effects could be sustained without fatigue (or running out of population!), the customer base would grow to over 150 billion.

     Why does this happen? First, notice that the interaction of the effects means that new customers brought in by advertising prime the viral pump, so to speak. Those purchased customers add to the base that is spreading the message. Second, customers brought in by the continuing viral effect produce enhanced viral impact while they are still new.

     The results of this interaction of the effects is illustrated in the following figure. The parameter values used here are one third of those in the figure above. The four lines on the graph represent the growth realized when different combinations of the effects are allowed to interact.

Graph of three types of growth interacting. The parameters for all three effects are reduced by 2/3 from those in the above graph. In addition, a 3% churn rate is included. The dark blue line represents a combination of all three growth effects interacting, with which the customer base doubles in less than half the time of the above graph. The other three lines represent different combinations of two effects. The dramatic lesson here is the power of a continuing viral effect. Without it, the light blue line shows, broadcast ads and new-customer virality cannot sustain the company against even a modest churn rate.

     The parameter values for this graph are as follows:

• Geometric (broadcast ads): $467 dollars a day is spent with an effectiveness of one customer acquired per dollar.
• Exponential due to new customers only: Starting with 25,000 new customers, each new customer brings in 0.35 others (or every 100 new customers bring in 35 more) after 30 days.
• Exponential due to all customers: Every customer brings in 0.01 new customers (or every 100^th customer brings in one new one) after every 30 days.
• Churn: To balance the growth effects, 3% of the customer base is assumed to defect in every 30-day period.

     The behavior illustrated above can be modeled for different combinations of parameter values using the Growth Math tool developed for this purpose. The above graphs were generated by it. Still more sophisticated models of feasible growth scenarios can be achieved by including the effects of market fatigue so that growth cannot be shown to continue ad infinitum.